JP3237292B2 - Alignment method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an alignment method for aligning an alignment pattern of a mask of an exposure apparatus with an alignment pattern formed by a plurality of underlayers in a photosensitive step in a semiconductor device manufacturing process.

[0002]

2. Description of the Related Art In a conventional exposure apparatus, when an alignment pattern is formed on each of a plurality of layers, an alignment pattern formed on a predetermined one of the films is detected, and a mask is formed accordingly. Had been aligned. Therefore, when the patterns are stacked over a plurality of layers to form a pattern of, for example, four layers, for example, the pattern of the second layer is matched with the pattern of the first layer, and the pattern of the third layer is formed. Is the second
Is matched to the pattern of the layer. Further, the pattern of the fourth layer is matched to the pattern of the third layer. Such alignment is performed. Alternatively, an alignment method for matching the patterns of the second, third, and fourth layers with reference to the pattern of the first layer has been proposed.

[0003]

However, when the alignment pattern of the mask is aligned with the alignment pattern formed by the layer immediately below the layer on which the pattern is to be formed, for example, the alignment accuracy of the exposure apparatus is reduced by a. Then, the amount of misalignment between the alignment pattern of the mask for forming the pattern of the fourth layer and the alignment pattern formed of the third layer becomes the above alignment accuracy a. Therefore, the amount of misalignment between the pattern formed by the fourth layer and the pattern formed by the third layer is a described above.

The amount of misalignment between the alignment pattern forming the fourth layer and the alignment pattern formed in the second layer is determined by the difference between the alignment pattern formed in the second layer and the alignment pattern formed in the third layer. Is equal to the square root of the sum of the squares of the misalignment amount a of the mask and the misalignment amount a between the alignment pattern of the mask forming the fourth layer and the alignment pattern formed of the third layer.
^1/2 a.

Similarly, the amount of misalignment between the alignment pattern formed of the first layer and the alignment pattern of the mask forming the fourth layer is 3 ^1/2 a. Therefore, the alignment pattern of the mask forming the fourth layer cannot be adjusted to an equal misalignment amount with respect to the alignment patterns formed of the first, second, and third layers.

Also, when the alignment patterns of the masks for forming the second, third and fourth layers are aligned with respect to the alignment pattern formed by the first layer, the alignment accuracy of the exposure apparatus is represented by a. Then, the second, third and fourth alignment patterns with respect to the alignment pattern formed by the first layer can be obtained.
The amount of misalignment of the alignment pattern of the mask of the layer is a.

Therefore, the amount of misalignment between the alignment pattern formed by the second layer and the alignment pattern formed by the third layer, or the alignment of the alignment pattern formed by the third layer with the mask forming the fourth layer. The amount of misalignment with the pattern is 2 ^1/2 a. For this reason, the alignment pattern of the mask forming the fourth layer cannot be adjusted to an equal misalignment amount with respect to the alignment patterns formed of the first, second, and third layers.

According to the present invention, for each of a plurality of underlying layers,
An object of the present invention is to provide an excellent alignment method for minimizing the amount of misalignment.

[0009]

SUMMARY OF THE INVENTION The present invention is an alignment method for achieving the above object. That is, the displacement position of each alignment pattern formed by the n-layer (n ≧ 2) underlayer is detected, and the position of the virtual alignment pattern is determined based on the displacement position of each detected alignment pattern to be aligned. I do. And an alignment method for aligning the mask alignment pattern with the virtual alignment pattern , wherein n
= 2, upper layer relative to lower alignment pattern
B, the alignment accuracy of the alignment pattern
Alignment of the mask with respect to the alignment pattern
Alignment accuracy of the lower and upper layers
Alignment pattern of the mask for the alignment pattern
Alignment accuracy c and down within [a ² + (b / 2) ^{2] 1/2}
And if n ≧ 3, the lowermost layer
Alignment pattern of intermediate layer for alignment pattern
Alignment pattern of the alignment accuracy and the intermediate layer over emissions
Alignment accuracy of the top layer alignment pattern
a for the alignment pattern of the top and bottom layers
The alignment accuracy c of the alignment pattern of the mask
Alignment to become [(n + 3) ^1/2 / 2] a
I do.

Alternatively, the virtual alignment pattern
The position overlaps the displacement position of the detected alignment pattern.
Brazing coefficientκ _A (However, 0 ≦ κ _A ≦ 1)Multiplied by
Coefficientκ _A Is determined based on each displacement position multiplied byBefore
Alignment of the mask with the virtual alignment pattern
An alignment method for matching the
= 2, upper layer relative to lower alignment pattern
B, the alignment accuracy of the alignment pattern
Alignment of the mask with respect to the alignment pattern
The alignment accuracy of the lower layer,
Alignment of the mask alignment pattern with the
Alignment accuracy c is [a ^Two + (Κ _A b) ^Two ] ^1/2 Become the upper layer
Alignment of Mask with Alignment Pattern
The pattern matching accuracy c is [a ^Two + (1-κ _A )
^Two b ^Two ] ^1/2 Aligned so that n ≧ 3
In the case, the middle layer
Alignment accuracy of alignment pattern and alignment of intermediate layer
Top alignment pattern for the alignment pattern
Alignment accuracy of the lowermost layer
Of mask alignment pattern
The degree c is [1+ (n-1) κ _A ^Two ] ^1/2 a
Alignment of mask to layer alignment pattern
The pattern matching accuracy c is [1+ (n−1) (1-κ)
_A ) ^Two ] ^1/2 Align so that it becomes a.

[0011]

[Action] In the above-described alignment method, by matching alignment pattern of the mask to the virtual alignment pattern determined based on the displacement position of each alignment pattern formed in the underlying layer of n layers, when n = 2, the lower
Mask alignment for layer and upper layer alignment pattern
The alignment precision c with the alignment pattern is [a ² + (b /
2) ² ] Do you align to ^1/2 ?
And if n ≧ 3 , alignment of top and bottom layers
Alignment of the mask alignment pattern with the pattern
So that the precision c becomes [(n + 3) ^1/2 / 2] a.
Because of the alignment, the mask is aligned with high precision for each underlying layer.

Further, the virtual alignment pattern is
By multiplying the detected displacement position of each alignment pattern by a weighting coefficient and determining based on the weighted coefficient, according to the allowable range of misalignment between the alignment pattern of the mask and the alignment pattern formed by each underlayer,
Alignment can be performed while changing the amount of misalignment between the mask and the pattern formed by each underlying layer. That is, the pattern of the mask is more precisely aligned with respect to the pattern formed with the base layer requiring high alignment accuracy than with the pattern formed with the other base layer.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention will be described with reference to FIG. In FIG. 1 , as an example,
A description will be given of a case where a mask pattern is aligned with a pattern formed by each of two underlying layers (n = 2) .

As shown in FIG. 1 , an alignment pattern 11 consisting of a first underlayer and an alignment pattern 21 consisting of a second underlayer above it are formed.
First, displacement positions of the alignment patterns 11 and 21 are detected. The displacement position is, for example, a position shifted from the design position. Now, the coordinates of the detected displacement positions of the alignment patterns 11 and 21 are represented by [A],
[B]. Such an alignment pattern 11,
In order to align the alignment pattern 31 of the mask mounted on the exposure apparatus (not shown) with the virtual pattern 21, it is assumed that a virtual alignment pattern 41 exists. And
The virtual alignment pattern 41 is set at a position having a uniform misalignment amount with respect to the alignment patterns 11 and 21. That is, on the pattern layout, a virtual alignment pattern 41 is set between the alignment pattern 11 and the alignment pattern 21. Then, the mask alignment pattern 31 is aligned with the virtual alignment pattern 41.

Here, the mask alignment pattern 3
1 and the misalignment accuracy between the alignment patterns 11 and 21 are obtained. Assume that the alignment accuracy of the alignment pattern 21 with respect to the alignment pattern 11 is, for example, b. The alignment accuracy of the mask alignment pattern 31 with respect to an arbitrary alignment pattern is represented by a.

Since the alignment accuracy between the alignment patterns 11 and 21 is b, the alignment accuracy between the alignment pattern 11 and the virtual alignment pattern 41 is b / 2. Similarly, the alignment accuracy between the alignment pattern 21 and the virtual alignment pattern 41 is b / 2. The position coordinates of the virtual alignment pattern 41 are [(A +) because the position coordinates of the alignment pattern 11 are [A] and the position coordinates of the alignment pattern 21 are [B].
B) / 2].

Therefore, the alignment accuracy between the alignment pattern 11 and the mask alignment pattern 31 is [a ²
+ (B / 2) ² ] ^1/2 . The alignment accuracy between the alignment pattern 21 and the mask alignment pattern 31 is also [a ² + (b / 2) ² ] ^1/2 . As described above, the alignment accuracy c between the alignment pattern 11 or the alignment pattern 21 and the mask alignment pattern 31 is c = [a ² + (b / 2) ² ] ^1/2 .

Usually, the exposure process is performed by the same or the same type of exposure apparatus.
If the alignment precision between b and a is equal to b, then the alignment patterns 11 and 21 and the mask alignment pattern 3
The matching accuracy c with 1 is c = (5 ^1/2 / 2) a.

In the alignment method described with reference to FIG. 1, an alignment pattern 1 formed by two underlayers is used.
By aligning the mask alignment pattern 31 with the displacement position of the virtual alignment pattern 41 obtained based on the displacement positions A and B of the masks 21 and 21, the alignment of the mask with respect to the alignment patterns 11 and 21 formed by the respective underlying layers The alignment pattern 31 has a positioning accuracy c = [a
² + (b / 2) ² ] ^1/2 .

Although the first embodiment has been described with reference to the case where the alignment pattern is formed with an alignment pattern formed of two underlayers, the same applies to the case of alignment with an alignment pattern formed of three or more base layers. Thus, the mask can be aligned. Next
Next, each of the alignment patterns is formed with n underlayers (n ≧ 3).
The case where is formed will be described with reference to FIG. In FIG.
Corresponds to the alignment method described with reference to FIG.
And a third underlayer (intermediate layer) between the first and second underlayers.
Is shown (n = 3).

As shown in FIG. 2 , an alignment pattern 11 consisting of a first underlayer and an alignment pattern 21 consisting of a second underlayer above the first underlayer.
And a third underlayer provided between the first and second underlayers
(Intermediate layer) . Therefore, the above alignment pattern 1
The alignment pattern 51 is aligned with the alignment pattern 51, and the alignment pattern 21 is aligned with the alignment pattern 51.

Assume that the alignment accuracy of the alignment pattern 51 with respect to the alignment pattern 11 is a, and
If the alignment accuracy of 1 is a, the alignment pattern 1
The alignment precision b of the alignment pattern 21 with respect to 1 is b = 2 ^1/2 a. Therefore, the mask alignment pattern 31 for the alignment patterns 11 and 21 is used.
Is equal to c = [a ² + (b / 2) ² ] ^1/2
= (6 ^1/2 / 2) a.

Although not shown, similarly, in the case of four layers,
Since b = 3 ^1/2 a, alignment pattern 1
The alignment accuracy c of the mask alignment pattern 31 with respect to 1, 21 is c = (7 ^1/2 / 2) a. Therefore, in the case of n layers, b = (n−1) ^1/2 a, so that the alignment accuracy c of the mask alignment pattern 31 with respect to the alignment patterns 11 and 21 is c =
[(N + 3) ^1/2 / 2] a.

Incidentally, in the conventional alignment method,
Assuming that the alignment accuracy in one alignment is a, for example, when the underlying layer is n layers (n ≧ 2), the alignment accuracy c3 of the mask alignment pattern 31 with respect to the alignment pattern 11 is c3 = (n) ^1/2 a become. Therefore, in the alignment method of the first embodiment, the alignment accuracy c1 of the mask alignment pattern 31 with respect to the alignment pattern 11 is c1 =
[(N + 3) ^1/2 / 2] a, so c3-c1 = n
It becomes ^{[(n + 3) 1/2} /2 ] ^{a> 0 - 1/2 a} . Therefore, in the above-described alignment method, the alignment accuracy is improved by [n ^1/2- (n + 3) ^1/2 / 2] a.

Next , as a second embodiment of the present invention, a case in which the displacement position of each alignment pattern is weighted will be described with reference to the explanatory diagram of the alignment method shown in FIG. Here, as an example, a case will be described in which the alignment pattern of the mask is adjusted with respect to the base layer composed of two layers. The same components as those described in FIG. 1 are denoted by the same reference numerals.

As shown in FIG. 3 , an alignment pattern 11 consisting of a first underlayer and an alignment pattern 21 consisting of a second underlayer above it are formed.
First, displacement positions of the alignment patterns 11 and 21 are detected. The displacement position is indicated by, for example, a deviation from a design position. Now, the coordinates of the detected displacement positions of the alignment patterns 11 and 21 are represented by [A] and [B].
And

Such alignment patterns 11 and 12
In order to match the alignment pattern 31 of the mask mounted on the exposure apparatus (not shown) with respect to 1, it is assumed that a virtual alignment pattern 41 exists. Then, the virtual alignment pattern 41 is weighted with respect to the alignment patterns 11 and 21 by the weighting factors κ _A and κ _B , respectively.
Is set to a position having an amount of misalignment multiplied by. Then, the mask alignment pattern 31 is aligned with the virtual alignment pattern 41.

Here, the mask alignment pattern 3
1 and the misalignment accuracy between the alignment patterns 11 and 21 are obtained. Assume that the alignment accuracy of the alignment pattern 21 with respect to the alignment pattern 11 is, for example, b. The alignment accuracy of the mask alignment pattern 31 with respect to an arbitrary alignment pattern is represented by a.

Now, since the alignment accuracy between the alignment patterns 11 and 21 is b, the alignment accuracy between the alignment pattern 11 and the virtual alignment pattern 41 can be obtained by setting the weighting coefficient to κ _A , where 0 ≦ κ _A ≦ 1. κ _A
b. Similarly, the alignment accuracy between the alignment pattern 21 and the virtual alignment pattern 41 is as follows: if the weighting coefficient is κ _B = 1−κ _A , where 0 ≦ κ _B ≦ 1,
It becomes _{κ B b = (1-κ} A) b. The position coordinates of the virtual alignment pattern 41 are the alignment patterns 1
Since the position coordinate of No. 1 is [A] and the position coordinate of alignment pattern 21 is [B], [κ _A A + (1−κ _A ) B]
become.

That is, when κ _A = 1, the alignment target of the mask is only the alignment pattern 11, and κ A
_{As A} becomes smaller than 1, the alignment target of the mask approaches the alignment pattern 21 and κ _A = 0.
In this case, the alignment target of the mask is only the alignment pattern 21. Thus, the weighting factors κ _A , κ
By multiplying by _B , the position coordinates of the virtual alignment pattern 41 are determined.

[0031] Therefore, if the alignment accuracy between the alignment pattern 11 and the virtual alignment pattern 41 as a, alignment accuracy between the alignment pattern 11 and the alignment pattern 31 of the mask is [a ² + (κ _A b)
² ] ^1/2 . The alignment accuracy between the alignment pattern 21 and the mask alignment pattern 31 is [a
² + becomes ^{_{(1-κ A) 2 b}} 2 ] ^1/2.

Usually, the exposure step is performed by the same or the same type of exposure apparatus.
If the alignment accuracy between 1 and 21 is a, the alignment accuracy c1 between the alignment pattern 11 and the mask alignment pattern 31 is c1 = (1 + κ _A ² ) ^1/2 a.
On the other hand, the alignment accuracy c2 between the alignment pattern 21 and the mask alignment pattern 31 is c2 = [1+
(1−κ _A ) ² ] ^1/2 a.

In the example described above, the displacement positions A and B of the alignment patterns 11 and 21 formed by a plurality of underlayers are described.
Is multiplied by the weighting coefficients κ _A and κ _B and the mask alignment pattern 31 is adjusted to the displacement position of the virtual alignment pattern 41 obtained based on the weighting coefficients κ _A and κ _B. , 21 according to the allowable range of the amount of misalignment between the mask and the pattern formed by each of the underlying layers. That is, the pattern of the mask is more precisely aligned with respect to the pattern formed with the base layer requiring high alignment accuracy than with the pattern formed with the other base layer.

In the second embodiment, similarly to the first embodiment, when the underlying layer is, for example, three or more layers, the mask alignment pattern 3
1 can be aligned. Next, the n layer
Each alignment pattern is formed on the underlayer of (n ≧ 3).
The displacement position of each alignment pattern.
The case where weighting is performed will be described with reference to FIG. In FIG.
Corresponds to the alignment method described with reference to FIG.
And a third underlayer (intermediate layer) between the first and second underlayers.
Is shown (n = 3).

As shown in FIG . 4 , an alignment pattern 11 made of a first underlayer and an alignment pattern 21 made of a second underlayer above the first underlayer.
And an alignment pattern 51 composed of a third underlayer provided between the first and second underlayers. Therefore, the alignment pattern 51 is aligned with the alignment pattern 11, and the alignment pattern 21 is aligned with the alignment pattern 51.

It is assumed that the alignment accuracy of the alignment pattern 51 with respect to the alignment pattern 11 is a, and the alignment pattern 2 with respect to the alignment pattern 51 is
If the alignment accuracy of 1 is a, the alignment pattern 1
The alignment precision “b” of the alignment pattern 21 with respect to 1 is b = ¹ 1a. The alignment accuracy between the alignment pattern 11 and the virtual alignment pattern 41, a weighting factor kappa _A, provided that if 0 ≦ κ _A ≦ 1 and, kappa _A
b. Similarly, the alignment accuracy between the alignment pattern 21 and the virtual alignment pattern 41 is as follows: if the weighting coefficient is κ _B = 1−κ _A , where 0 ≦ κ _B ≦ 1,
It becomes _{κ B b = (1-κ} A) b. Therefore, the alignment pattern 3 of the mask with respect to the alignment pattern 11
The alignment accuracy c1 of 1 is c1 = (1 + 2κ _A ² ) ^1/2
a. The alignment accuracy c2 of the mask alignment pattern 31 with respect to the alignment pattern 21 is c2 =
[1 + 2 (1-κ _A ) ² ] ^1/2 a.

Although not shown, similarly, in the case of four layers,
Since b = 3 ^1/2 a, c1 = (1 + 3κ _A ² )
^1/2 a. Also, c2 = [1 + 3 (1-κ _A ) ² ]
^1/2 a. Therefore, in the case of n layers, b =
Since (n−1) ^1/2 a, the alignment accuracy c of the mask alignment pattern 31 with respect to the alignment pattern 11 is c1 = [1+ (n−1) κ _A ² ] ^1/2 a
And the alignment accuracy c of the mask alignment pattern 31 with respect to the alignment pattern 21 is c2 = [1
+ (N-1) (1- [kappa] _A ) ² ] ^{1 /} 2a.

[0038]

As described above, according to the present invention,
Since the alignment pattern of the mask is aligned with the virtual alignment pattern obtained based on the displacement position of each alignment pattern formed by the n base layers , n =
In the case of 2, for the alignment pattern of the lower layer and the upper layer
The alignment accuracy c with the mask alignment pattern is [a
² + (b / 2) ² ] aligned to ^1/2
And if n ≧ 3, alignment of the top and bottom layers
Alignment of the mask alignment pattern with the pattern
So that the precision c becomes [(n + 3) ^1/2 / 2] a.
Since the alignment is performed , the mask can be positioned with high accuracy with respect to each underlying layer.

Further, the virtual alignment pattern is
Since the detected displacement position of each alignment pattern is multiplied by a weighting coefficient and obtained based on the weighted coefficient, the mask and each underlying layer are adjusted according to the allowable range of misalignment between the alignment pattern of the mask and the alignment pattern formed by each underlying layer. It is possible to change the amount of misalignment with the pattern formed by the method. For this reason, the pattern of the mask can be more precisely aligned with respect to the pattern formed with the base layer requiring high alignment accuracy than with the pattern formed with the other base layer.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram of an alignment method according to a first embodiment.

FIG. 2 is an explanatory diagram of alignment accuracy in the case of three underlayers.

FIG. 3 is an explanatory diagram of an alignment method according to a second embodiment.

FIG. 4 is an explanatory diagram of alignment accuracy in the case of three underlayers.

[Explanation of symbols]

11 : alignment pattern , 21 : alignment pattern , 31 : mask alignment pattern , 41 : virtual alignment pattern , κ _A : weighting coefficient , κ _B ...
Weighting factor

Claims (2)

(57) [Claims] 1. An alignment method for aligning an alignment pattern of a mask of an exposure apparatus with an alignment pattern formed of an n-layer (n ≧ 2) underlayer, wherein a displacement of each alignment pattern formed by the n-layer underlayer is provided. An alignment method for detecting a position, determining a position of a virtual alignment pattern based on a displacement position of each of the detected alignment patterns to be aligned, and then aligning the mask alignment pattern with the virtual alignment pattern. When n = 2, the upper position with respect to the lower alignment pattern
The alignment accuracy of the layer alignment pattern is b,
Alignment of the mask with respect to the alignment pattern of
Assuming that the alignment accuracy of the mask pattern is a, the mask
The alignment accuracy c with the alignment pattern is [a ² + (b /
2) ² ] Alignment so as to be ^1/2 , and when n ≧ 3, the alignment with respect to the bottommost alignment pattern
Alignment accuracy of intermediate layer alignment pattern and intermediate
Top layer alignment to layer alignment pattern
The combined accuracy of the preparative pattern as a, a mask for the top and bottom layers of the alignment pattern
The alignment accuracy c of the alignment pattern is [(n + 3)
^[ 1/2] / 2] a . 2. The alignment method according to claim 1, wherein a displacement position of each of the alignment patterns formed in the n underlayers is detected, and the detected displacement position of each of the alignment patterns is multiplied by a weighting coefficient to weight. after determining the position of the virtual alignment pattern on the basis of the respective displacement position multiplied by the coefficient, a alignment method to align the alignment pattern of the mask to the virtual alignment pattern, if n = 2, the lower layer of the alignment pattern On
The alignment accuracy of the layer alignment pattern is b,
Alignment of the mask with respect to the alignment pattern of
A, and the weighting coefficient is κ.
_A (where 0 ≦ κ _A ≦ 1), the alignment of the mask to the lower alignment pattern
Alignment accuracy c of cement pattern _{^{[a 2 + (κ A b)}} 2 ]
^{1/2 and} the alignment pattern for the upper layer
The alignment accuracy c of the mask alignment pattern is [a ² +
(1-κ _A ) ² b ² ] Aligned to ^1/2
When n ≧ 3, the alignment pattern of the lowermost layer is
Alignment accuracy of intermediate layer alignment pattern and intermediate
Top layer alignment to layer alignment pattern
_{A, and the} weighting coefficient is κ _A
(However, 0 ≦ κ _A ≦ 1) and the alignment of the mask with the lowermost alignment pattern
The alignment precision c of the alignment pattern is [1+ (n-1) κ
_A ² ] ^{1 /} 2a and the topmost alignment pattern
Accuracy c of the mask alignment pattern with respect to
Becomes [1+ (n-1) (1-κ _A ) ² ] ^1/2 a
Alignment method which is characterized in that sea urchin alignment.